Subscriber access provided by University of Sydney Library
Review
Analytical Aspects of Nanotoxicology Ian L. Gunsolus, and Christy L. Haynes Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b04221 • Publication Date (Web): 12 Nov 2015 Downloaded from http://pubs.acs.org on November 15, 2015
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Title: Analytical Aspects of Nanotoxicology Authors: Ian L. Gunsolusa and Christy L. Haynesa a
Department of Chemistry, University of Minnesota, Minneapolis, MN, 55455
Corresponding Author: Christy L. Haynes; 207 Pleasant St. SE, Minneapolis, MN, 55455; phone: 612-6261096;
[email protected] Author Biographies: Ian Gunsolus is a doctoral candidate in Chemistry at the University of Minnesota under the direction of Christy Haynes. He received a B.A. in Chemistry and a B.A. in Mathematics from St. Olaf College in 2011, and a M.S. in Chemistry from the University of Minnesota in 2013. His current research focuses on the environmental toxicity of engineered nanomaterials. Christy Haynes is the Elmore H. Northey Professor of Chemistry at the University of Minnesota where she leads the Haynes Research Group, a lab dedicated to applying analytical chemistry to the fields of immunology and toxicology. Professor Haynes completed her undergraduate work at Macalester College in 1998 and earned a Ph.D. in chemistry at Northwestern University in 2003 under the direction of Richard P. Van Duyne. Before joining the faculty at the University of Minnesota in 2005, Haynes performed postdoctoral research in the laboratory of R. Mark Wightman at the University of North Carolina, Chapel Hill. Among many honors, she was named a 2010 Alfred P. Sloan fellow and a National Institutes of Health "New Innovator." She was also recently honored by the Royal Society of Chemistry with the Joseph Black Award, which recognizes achievement by young scientists, and by the American Chemical Society Division of Analytical Chemistry with the Arthur F. Findeis Award for Achievements by a Young Analytical Scientist.
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Introduction Nanotoxicology, the study of nanomaterial toxicity to biological systems, has become a vibrant area of research over the past ten years. Two main factors drive this enthusiasm: industrial interest in engineered nanomaterials,1 leading to their increasing use in consumer products2,3 and increasing possibility of human and environmental exposure;4 and the unique physical and chemical properties of nanomaterials relative to their bulk counterparts. The rapid growth of academic research in nanotoxicology is evidenced by the sharp increase in the number of articles published on this subject over the past approximately ten years (Figure 1). This review highlights the analytical methods used in nanotoxicology research, encompassing methods that probe the dynamic physical and chemical properties of nanomaterials (section 1) as well as methods that assess interactions between nanomaterials and biological systems (section 2). We will restrict our discussion to the toxicity of engineered nanomaterials, the subjects of the vast majority of nanotoxicity studies to date, as opposed to naturally occurring nanomaterials. The content of this review will be weighted towards work that has been published within the past three years; almost 50% of our citations fall in the period 2013-2015.
Figure 1. Returns for the search term “nanotoxicology” on SciFinder as of Oct. 14, 2015, organized by publication year.
Unique experimental challenges distinguish nanotoxicology research from the more classical field of molecular toxicology. Among these are the highly dynamic chemical composition and physical properties of nanomaterials (Figure 2). Nanomaterials routinely undergo both chemical and physical transformations even under controlled laboratory conditions through mechanisms including dissolution,
ACS Paragon Plus Environment
Page 2 of 55
Page 3 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
release of ligands that decorate their surface, adsorption of secondary species (e.g., proteins), and aggregation. Characterizing these dynamic chemical and physical properties is crucial to identifying the mechanism of interaction with biological systems and the nanomaterial toxicity. Even in the absence of these transformations, a population of nanomaterials does not have a uniform chemical or physical identity. While a mole of a particular molecule is composed of chemically identical units and is thus chemically uniform, a mole of a particular nanomaterial exhibits a distribution of chemical and physical parameters like oxidation state, size, and shape.
Figure 2. Physical and chemical transformations of silver nanoparticles are depicted to illustrate the dynamic properties of nanomaterials in general. Arrows are used to guide the eye along the progression of a particular transformation over time. Depicted from left to right: agglomeration of protein-coated nanoparticles (containing a diffuse layer around the nanoparticle core); oxidation of nanoparticles to release ionic silver; oxidation of the nanoparticle surface to form a layer of silver oxide; and agglomeration of pristine nanoparticles.
An additional, and equally important, set of experimental challenges arises from the great variability in chemical composition and structural complexity of commercially used nanomaterials. In commercial products, nanomaterials are used as chemically inert or active additives that impart desired qualities, such as increased hardness or surface area, antimicrobial behavior, UV protection, and coloring.5 In some important emerging applications, nanomaterials serve not just as useful additives, but form the basis of product performance. This includes drug-delivery for cancer therapy,6 energy conversion7 and energy storage materials.8 Nanomaterials used in these and other commercial applications are highly variable in size, shape, chemical composition, surface functionalization, crystallinity, and incorporation
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
into secondary structures (e.g., polymer films). The chemical and structural diversity of nanomaterials, which is likely to increase as new generations of nanomaterials are developed, leads to highly variable modes of interaction with biological systems, challenging the establishment of standard toxicological methods. This challenge has been addressed in two primary ways in the nanotoxicology literature. In the first approach, known routes of nanotoxicity, including disruption of cell membrane integrity, DNA damage, and oxidative stress, are assessed. Commercially available cytotoxicity assays, primarily based on small molecule probes of a particular biological pathway or endpoint, have been frequently employed for this purpose. However, some assays experience significant interference resulting from probe interactions with nanomaterials;9 implicated nanomaterials include silver, titania, copper, and iron oxide nanoparticles as well as carbon nanotubes10–14. While these assays provide extensive insights into the toxicological effects of nanomaterials on biological systems, their specificity dictates that only a small set of hypotheses related to the nanomaterial-biological interaction can be tested. Methods with more general analytical capabilities are also needed to enable hypothesis-driven investigations of nanomaterial-biological interactions that are not limited to already known routes of toxicity. Such methods are more likely to reveal unknown or unexpected mechanisms of nanomaterial-biological interaction, a critical step towards predicting nanomaterial risk and developing sustainable nanomaterials. In a second approach, nanotoxicologists are increasingly adapting analytical methods with origins in the chemical sciences (e.g., spectroscopy, mass spectrometry, and electrochemical methods) and the biological sciences (e.g., genomics, proteomics, metabolomics, and microscopy) to assess the response of biological systems to nanomaterial exposure. This review focuses on the application of analytical methods in two broad categories. The first is methods used to characterize the physical interaction of nanomaterials with biological systems based on uptake and distribution in cells and tissues and biophysical changes in cell membrane integrity, and DNA and protein structure. The second is methods used to characterize changes in biological function following exposure to nanomaterials. These include changes in cell and organism survival and reproduction; mobility; signaling; and gene, protein, and small molecule expression. Continuing to develop analytical methodology that probes the dynamic interface between nanomaterials and biological systems, rather than just the toxicological outcome of this interaction, remains a fundamental challenge in the field of nanotoxicology, but will enable more causal, as opposed to correlative, mechanisms of nanotoxicity to be discovered.15 We suggest that causal, mechanistic knowledge of nanotoxicity is necessary to design functional nanomaterials with controlled biological impacts and to keep pace with the rate of new and increasingly complex nanomaterial incorporation into consumer products. Section 1 of this review will focus nanomaterials themselves, reviewing methods that facilitate characterization of nanomaterial aggregation, degradation, and molecular surface adsorption. 1. Characterizing Dynamic Physical and Chemical Properties of Nanomaterials 1.1 Aggregation and Degradation Microscopy. Transmission electron microscopy (TEM) remains the gold-standard for characterizing the primary size and structure of nanomaterials based on its high spatial resolution. Chemical information
ACS Paragon Plus Environment
Page 4 of 55
Page 5 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
about a nanomaterial can also be obtained through parallel analysis using spectroscopic methods such as electron energy loss spectroscopy (EELS) or energy-dispersive x-ray spectroscopy (EDX). Despite this versatility, sample preparation requirements for TEM traditionally limits its utility for characterizing nanomaterial dispersions in situ (i.e., under biologically or environmentally relevant conditions). In particular, the high-vacuum environment of electron microscopes can lead to sample preparation artifacts (e.g., nanomaterial aggregation induced by solvent evaporation) that do not reflect nanomaterial behavior in solution. However, advances in sample preparation are expanding the utility of electron microscopy for in situ nanomaterial characterization. Cryo-TEM, in which samples are imaged in a vitrified state, preserves the spatial arrangement of species as they exist in solution. This technique has been used, for example, to observe decreased packing density of poly(organosiloxane) nanoparticles in the presence of serum, indicating possible co-aggregation of nanoparticles and serum proteins.16 Cryo-TEM has also been used probe the molecular structure of hybrid perylene diimide/single wall carbon nanotube structures.17 Based on the observed distance between adjacent carbon nanotubes in cryo-TEM images, the authors concluded that a single molecular layer of perylene diimide derivative exfoliated the nanotubes. While cryo-TEM is a powerful probe of nanoscale packing, it cannot asses the dynamic behavior of nanomaterial dispersions. Liquid-cell TEM does preserve the fluid environment around the nanomaterial. In this technique, imaging electrons pass through a viewing window (typically composed of Si3N4) and encounter a liquid sample sealed in a chamber. Liquid-cell TEM has been used extensively to observe the crystal growth of metal nanoparticles, the subject of a recent review,18 and liquid-cell scanning TEM has been recently used to observe the hydration layer around alumina nanoparticles, representing the first time that the solvent shell (i.e., the region of solvent which is restructured relative to the bulk solvent through its interaction with the nanomaterial surface19) has been directly visualized.20 Recent development of a highly electron-transparent liquid cell made from graphene sheets has enabled liquid-cell TEM to observe the molecular motion of nano-bio conjugates (composed of Au nanoparticles and doublestranded DNA) in solution.21 While “soft” biological structures such as DNA are susceptible to damage from prolonged exposure to an imaging electron beam, liquid-cell TEM offers a means to observe dynamic behavior of nano-bio structures with very high spatial resolution. Despite these advances, long image acquisition and analysis times make these methods ill-suited for high-throughput, statistically significant characterization of nanomaterial properties. Light microscopy is a less-invasive, simpler, and faster alternative to electron microscopy, but is generally unsuitable for characterizing nanomaterial dispersions due to its much lower spatial resolution. However, dark-field microscopy combined with hyperspectral imaging (e.g., the CytoViva® Hyperspectral microscope) is an emerging light microscopy method suitable for nanomaterial characterization. In the CytoViva® imaging system, a novel dark-field condenser illuminates a sample with highly collimated light at oblique angles, improving signal-to-noise and contrast over traditional dark field microscopy.22 Scattered light in the range 400-1000 nm is resolved using a microscopemounted spectrophotometer and spectrophotometer-integrated CCD and analyzed using environment for visualizing images (ENVI) software. Badireddy et al. employed the CytoViva® imaging system to differentiate between silver, titania, and ceria nanoparticles in a mixed sample in pure water, and to detect and quantify the relative amount of silver nanoparticles in complex aqueous samples containing natural colloidal species (i.e., simulated wetland ecosystem water and clarified wastewater).23 Engineered nanoparticles were detected and classified by spectral similarity analysis; spectra collected
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
from a sample were compared to a library of reference spectra collected for each nanoparticle in pure water, and matches were determined by setting an arbitrary similarity threshold of 90%. In another study, Badireddy et al. used similar methods to observe silver nanoparticle formation from silver ions in solutions containing chloride or bacterial cells and extracellular polymeric substances exposed to visible light (Figure 3).24 Microscopy is best used in combination with ensemble measurements that can analyze a much larger population of nanomaterials while also characterizing their dynamic, hydrated behavior. Nanomaterials are prone to aggregation in high ionic strength solutions, including many with biological and environmental relevance (e.g., blood plasma and saltwater). Repulsive electrical double layer interactions between nanomaterials are reduced as ionic strength increases, and when attractive van der Waals forces dominate, aggregation occurs. Nanomaterials frequently exhibit time-variable mobility under the conditions of most toxicological assays, where aggregation increases the particle diffusion constant and can lead to particle settling. Dynamic changes in aggregation state and mobility can significantly affect nanomaterial toxicity by influencing the nanomaterial dose and effective size that comes in direct contact with a biological system. Consequently, it is imperative that nanomaterial aggregation and mobility be understood in order to identify the source of nanomaterial toxicity. There are a variety of analytical methods available to characterize the dynamic, hydrated behavior of nanomaterials.
Figure 3. Darkfield hyperspectral images of suspected individual Ag nanoparticles are shown in panels ac and e-g (400x magnification, scale bar not provided). Images were collected from a suspension of AgCl exposed to visible light (a, e) and a suspension of the bacterium Pseudomonas aeruginosa in medium containing NaCl and extracellular polymeric substances (b, f). Visible near-infrared spectra (d, h) collected from these suspected nanoparticles from the areas highlighted in crosshairs closely match those collected from polyvinylpyrrolidone-stabilized Ag nanoparticles prepared directly (c, g), suggesting nanoparticle formation under the conditions studied. Reprinted with permission from Badireddy, A. R.; Budarz, J. F.; Marinakos, S. M.; Chellam, S.; Wiesner, M. R. Environ. Eng. Sci. 2014, 31, 338–349 (ref 24). The publisher for this copyrighted material is Mary Ann Liebert, Inc. publishers.
ACS Paragon Plus Environment
Page 6 of 55
Page 7 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Dynamic light scattering (DLS) is the most commonly used method to determine nanomaterial mobility, and is based on measurements of hydrated diameter in solution. Dynamic fluctuations in the intensity of light scattered by nanomaterials (due to Brownian motion) are monitored and fit to an autocorrelation function. From this fit, a nanomaterial’s diffusion coefficient can be obtained and subsequent application of the Stokes-Einstein relation allows the hydrodynamic diameter to be calculated. Although DLS is most commonly used to probe static nanomaterial suspensions, it can also operate under continuous flow as, for example, when coupled with asymmetric field flow fractionation (vide infra) to simultaneously fractionate and characterize nanoparticle dispersions.25 DLS assumes a spherical particle geometry and thus has limited utility for characterizing the behavior of non-spherical nanomaterials (e.g., commonly used materials such as gold nanorods and carbon nanotubes). In addition, DLS has greater sensitivity for larger particles than smaller ones,26 decreasing its ability to resolve multiple particle size distributions. Light scattering from other colloidal species present in many biologically and environmentally relevant matrices (e.g., protein aggregates and dissolved organic matter) can interfere with DLS measurements of nanomaterial behavior since, like most methods used to track nanomaterial aggregation and mobility, DLS lacks chemical selectivity.27 Nanoparticle tracking analysis (NTA) has better resolution than DLS for characterizing mobility and size of nanoparticles in dispersions. Like DLS, NTA detects scattered light incident on the particles; unlike DLS, it uses a high-frame-rate digital camera to acquire images of nanoparticles in solution. The Brownian motion of individual particles is observed by tracking the evolution of their position across many images, and the Stokes-Einstein relation is applied to calculate the hydrodynamic diameter. Since NTA can discriminate between individual particles, it is typically better suited to resolving distinct particle populations in polydisperse samples than is DLS, which acquires an ensemble measurement of all particles simultaneously. Images obtained from NTA can also be used to visualize the sample and determine the approximate nanoparticle concentration.28 NTA assumes, as does DLS, a spherical morphology to calculate particle size from diffusion rate, making it generally ill-suited to studying nonspherical nanomaterials. While the resolving power of NTA is generally considered superior to that of DLS, it is not adequate for all applications. For example, a study by Anderson et al. showed that NTA and DLS could not resolve populations of polystyrene nanoparticles in a dispersion containing 220- to 410nm diameter particles.29 Tunable resistive pulse sensing (TRPS) can be used when higher resolution is required to distinguish multiple particle populations in a polydisperse sample. The number density distribution of particle sizes obtainable with TRPS allows more accurate determination of particle mobility and morphology than is possible with DLS. This method relies on the Coulter principle, widely exploited in Coulter counters used to size and count cells.30 The Coulter principle states that the flow of a particle through a pore containing conductive fluid transiently increases the electrical impedance across the pore in a manner proportional to the particle’s volume.31 While traditional Coulter counters utilize rigid pores of a fixed size, TRPS employs pores fabricated in an elastic membrane. Pore size can be precisely controlled by changing the tension across the membrane; this increases the resolving power and dynamic range of TRPS.32 Pal et al. have used TRPS to individually measure the hydrated size and aggregation behavior of populations of multiple nanomaterials with greater sensitivity and resolution than is possible with DLS (Figure 4).33 Particles studied include non-spherical carbon-based nanomaterials, cerium oxide
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
nanoparticles, and nickel nanoparticles. TRPS can be used to simultaneously monitor nanoparticle zetapotential and size, based on a theoretical treatment of the forces acting on the particle which determine its velocity through a pore.34 Analysis of the impedance peak traces resulting from TRPS analyses (e.g., half-width) may reveal additional information regarding dynamic changes in particle charge and other parameters, such as protein corona thickness.33
Figure 4. Comparison of DLS- and TRPS-measured size-distributions of H2O2-oxidized single wall carbon nanohorns prepared in cell culture medium (RPMI medium) supplemented with 10% fetal bovine serum. Sequential dilutions of a 500 µg/mL stock solution of nanohorns in medium were prepared to achieve the concentrations listed. DLS analysis indicated a single, broad distribution of nanohorn sizes at most nanomaterial concentrations, though interfering signal in small size regimes due to serum proteins was observed at low nanohorn concentrations. In contrast, TRPS resolved two distinct nanohorn populations at all concentrations studied. Measured average nanohorn size varied with concentration when measured by DLS, while TRPS-measured averages showed limited concentration-dependence. Reprinted with permission from Pal, A. K.; Aalaei, I.; Gadde, S.; Gaines, P.; Schmidt, D.; Demokritou, P.; Bello, D. ACS Nano 2014, 8 (9), 9003–9015 (ref 33). Copyright 2014 American Chemical Society. Electrospray-differential mobility analysis (ES-DMA) has been used to characterize the aggregation behavior of spherical and non-spherical nanomaterials, including silver35 and gold nanoparticles,27,36–38 gold nanorods,39 graphene nanosheets,40 and single-walled carbon nanotubes.41 This technique measures the electrical mobility of aerosolized nanoparticles in a sheath flow of air under an applied electric field of variable strength. To assess the properties of non-spherical particles, including nanosheet thickness, nanorod diameter and length, and nanotube length, extensive data fitting and
ACS Paragon Plus Environment
Page 8 of 55
Page 9 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
correction parameters must be employed. These extensive data processing requirements limit the usefulness of ES-DMA in routine analysis of non-spherical nanomaterials. Single particle inductively coupled plasma mass spectrometry (sp-ICP-MS) addresses the lack of chemical selectivity inherent in other nanomaterial aggregation and mobility analysis methods. Most methods used to track nanomaterial aggregation and mobility cannot distinguish between nanomaterials of interest and other colloidal species common to many biological matrices (e.g., lipid vesicles, dissolved organic matter, or protein aggregates), potentially decreasing the accuracy of nanomaterial size and mobility measurements due to signal interference. The sp-ICP-MS method significantly expands the capability of traditional ICP-MS for studying nanomaterials by incorporating time-resolution into elemental composition analysis. Ionized colloidal particles produce transient signals in the MS detector, assuming an optimized dwell-time and sufficiently dilute particle concentration such that particle ionization and detection events do not overlap. Transient signals are analyzed according to their frequency, intensity, and elemental composition; from these data, colloidal concentration, size distribution, and chemical identity, respectively, can be calculated.42–45 Soluble species produce signals with more constant intensity, and their concentration can be separately quantified. The ability of sp-ICPMS to distinguish colloidal species from soluble ions makes it a valuable tool for studying the toxicity of nanomaterials prone to chemical degradation, as these ions can contribute significantly to overall toxicity.46–48 The sp-ICP-MS method has been shown to have a similar accuracy to DLS for determining nanoparticle size distribution.49 Peters et al. have recently demonstrated the robustness of sp-ICP-MS for nanoparticle detection in various complex matrices, including food, waste water, culture media, and biological tissue, using four commonly used nanoparticles (Au, Ag, TiO2, and SiO2).50 Tuoriniemi et al. reported a size detection limit of 20 nm for Ag nanoparticles; below this, signal intensity spikes due to nanoparticle ionization overlapped with the background signal of dissolved silver ions.51 Liu et al. demonstrated a size detection lower limit of 10 nm for gold nanoparticles. An upper limit of 70 nm was obtained when using a pulse mode counting detector.52 Above this nanoparticle size, the signal spike produced upon nanoparticle ionization exceeded the linear dynamic range of the detector. Operating in a lower sensitivity analogue mode allowed the authors to extend the size detection maximum to 200 nm. Liu et al.’s study also demonstrates the capability of sp-ICP-MS to simultaneously monitor nanomaterial size and chemical speciation. Dissolved Au ions released from gold nanoparticles contributed constant signal in the MS detector, while gold nanoparticles contributed spikes corresponding to individual nanoparticle events.52 Laborda et al. reported a similar ability to distinguish dissolved Ag ions and Ag nanoparticles.53 In parallel, the authors measured the size distribution and number concentration of silver nanoparticles. Nanoparticle sizing by sp-ICP-MS generally assumes a spherical particle morphology and compares the signal intensity from an unknown sample to a series of standards. Nanoparticle diameter is calculated using the geometric relation between spherical particle mass and diameter.54 To the best of our knowledge, sp-ICP-MS has not been applied to characterize non-spherical nanomaterials.
Differential centrifugal sedimentation is a very high-resolution method to determine nanoparticle size that relies on tracking nanoparticle transport through a density gradient.29 Nanoparticles are introduced into a carrier solution (typically sucrose), which establishes a density gradient as the solution spins and is subject to centrifugal force. While this method has high resolution and a large dynamic range, it cannot
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
be used for in situ measurement of nanoparticle transformations in biological matrices due to the requirement that nanoparticles be introduced into a carrier solution.29 However, a centrifugal method for particle size discrimination that can be applied to nanoparticles suspended in biologically relevant matrices has been developed by DeLoid et al.55 These researchers developed a standard method using volumetric centrifugation (pellet volume determination) to determine the effective density of nanomaterials, which they applied to calculate delivered nanomaterial dose to surface-bound cell cultures, a common mode of toxicological assesment.56 The latter calculation was achieved using the in vitro sedimentation, diffusion and dosimetry (ISDD) model created by Hinderliter et al, which uses nanomaterial hydrated size (measured by DLS) and density as inputs.57 Ion selective electrodes (ISEs) offer a means to monitor the chemical speciation/complexation of nanomaterials in complex matrices and are primarily useful for tracking the dissolution of chemically unstable nanomaterials. ISEs are potentiometric chemical sensors with high selectivity for a target ionic analyte and are routinely employed for in situ quantification in complex matrices such as plasma and whole blood.58 The earliest ISE, the glass pH electrode, was based on the selective permeability of glass for hydrogen ions. State-of-the-art ISEs use ion receptors (i.e., ionophores).59 ISEs employing ionophores present in matrices of highly fluorinated compounds have shown very high selectivities and stabilities, making them among the most selective chemical sensors.60–63 While ISEs are not capable of analyzing nanomaterial size in addition to chemical speciation, as is possible using sp-ICP-MS, their capacity to monitor chemical speciation with high selectivity and timeresolution is superior to that of sp-ICP-MS. Commercially available solid-state Ag/Ag2S ISEs have been used to monitor silver nanoparticle speciation (i.e., to observe release of Ag(I) from Ag nanoparticles) in a number of studies.64–66 Käkinen et al. monitored Cu (II) release from Cu nanoparticles in various media commonly used for biological toxicity testing using both a commercially available Cu-ISE and a bioluminescent bacterial-based (Pseudomonas fluorescens) biosensor.67 The authors observed that undissolved copper, inherently undetectable by the ISE, was detected by the bacterial biosensor, suggesting that undissolved metal from the nanoparticles was bioavailable. This study suggests that both dissolved metal ions, commonly implicated for metal nanomaterial toxicity, and undissolved nanomaterial may be bioavailable and contribute to overall organismal toxicity. Custom-made ISEs also have been used to monitor nanoparticle dissolution. Maurer-Jones et al. employed a fluorous-phase silver ISE to monitor silver nanoparticle dissolution in real-time in a biological matrix containing live bacterial cells (Shewanella oneidensis MR-1).68 They observed altered nanoparticle dissolution kinetics and lower total silver ion release from nanoparticles in the presence of cells than in an identical cell-free matrix, demonstrating that the presence of organisms can alter nanoparticle dissolution behavior. Koch et al. similarly observed decreased silver nanoparticle dissolution in the presence of human alveolar epithelial cells versus in cell-free culture media using a commercially available solid-state membrane Ag/Ag2S ISE.69 As discussed here, multiple analytical methods are available to characterize the dynamic, hydrated behavior of nanomaterials. Given the increasingly common occurrence of nanomaterials with nonspherical morphologies (e.g., nanosheets, nanotubes, nanocubes), it is imperative that researchers move beyond light scattering techniques that assume spherical geometry as the primary means to characterize nanomaterial aggregation behavior. Electron microscopy remains the most readily available option for many researchers to characterize both spherical and non-spherical nanomaterial aggregation
ACS Paragon Plus Environment
Page 10 of 55
Page 11 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
or size change due to dissolution, though its low throughput and small sampling size present significant limitations. TRPS and ES-DMA represent two high-throughput options to characterize both spherical and non-spherical nanomaterials in complex, biologically relevant matrices, and commercial instruments are available. TRPS will be much more easily implemented by most nanotoxicologists, as a result of far lower data processing requirements than ES-DMA. ISEs provide a means to monitor the chemical speciation of nanomaterials with exceptionally high selectivity while sp-ICP-MS, though its application to nanomaterials with non-spherical morphology remains unexplored, offers the unique capability to simultaneously monitor nanomaterial size and chemical speciation. 1.2 Nanomaterial Surface Adsorption Nanomaterials exhibit not only dynamic colloidal behavior under biologically relevant conditions, but dynamic changes in surface chemistry as well. Components of biological and environmental matrices, including proteins70,71 and lipoproteins72 and dissolved organic matter,73,74 can adsorb onto nanomaterial surfaces under realistic exposure scenarios, forming a molecular “corona”. Protein coronas, being the most widely studied, have been shown to significantly alter nanomaterial colloidal behavior and interaction with biological systems relative to the pristine nanomaterial by modifying factors such as nanomaterial size, surface charge, and surface chemistry.75 The composition of the protein corona, e.g. the glycosylation state of adsorbed proteins, can influence the extent of nanomaterial internalization by cells.76,77 Adsorbed proteins can also influence immune response in cells, a subject that has been recently reviewed.78 Adsorption of other molecules, including humic and fulvic acids, has also been shown to alter nanomaterial colloidal behavior and toxicity.79–81 Detailed characterization of the dynamic composition of the protein corona, as well as coronas formed from less-studied molecules like dissolved organic matter, is necessary to understand the interface between nanomaterials and biological systems under biologically and environmentally relevant exposure scenarios. The binding constants controlling molecular adsorption to nanomaterial surfaces and the composition, surface-coverage, thickness, and orientation of molecules on the nanomaterial surface have been studied using a number of analytical techniques. A significant barrier to characterization of molecular adsorption to nanomaterial surfaces in situ is signal interference due to the high chemical complexity of biologically relevant media (which may contain proteins, lipids, and dissolved organic matter). Additionally, discriminating between species adsorbed to the nanoparticle surface and those in solution can be difficult due to their similar chemical signatures. These problems can be ameliorated by isolating subpopulations of nanomaterials or by isolating nanomaterials from interfering colloidal species in the sample matrix prior to analysis; however, separation methods have been infrequently employed in published studies. Currently available separation methods include column chromatography and field flow fractionation and its derivatives (see Section 1.2.2). Development of more rapidly implementable and generalizable methods to separate nanomaterials from complex biological matrices would facilitate accurate nanomaterial characterization. 1.2.1 Probes of Molecules at Nanomaterial Surfaces: X-ray photoelectron spectroscopy (XPS) is widely used to probe the chemical environment of planar surfaces, but is more difficult to perform on nanoparticle surfaces due to their non-planar geometry and small size.82 Torelli et al. have recently developed a quantitative method to correct XPS data obtained from small, spherical, surface-functionalized nanoparticles by accounting for the effects of surface curvature on spectral intensities. Their geometric correction method can be applied to quantitatively
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
measure ligand density on the surface of nanoparticles smaller than 20 nm using XPS, which would otherwise give aberrant quantitative data.82 Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy is used to assess the adsorption of molecules to nanomaterial surfaces both qualitatively (identity and structure of surfaceadsorbed ligands) and quantitatively (kinetics of adsorption).83 ATR-FTIR can be used to study the interactions of both small molecules (e.g., carboxylic acids,84,85 and catechols86) and large molecules (e.g., proteins) with nanomaterials, though studies of larger molecules are less common. Nanomaterials are introduced to the ATR crystal surface by drop casting, which creates a thin film, and molecules of interest are then introduced using a flow-cell placed on the crystal’s surface. An evanescent infrared wave, produced as a result of total internal reflection at the crystal surface, serves as a probe of the chemical environment within a short (~ 1 µm) distance from the crystal-solution interface. This short penetration depth limits signal contributions from the bulk solution (i.e., from species not bound to the nanomaterial surface), although interfering signal from the bulk may still be significant enough to complicate analysis in complex solution environments. Qualitative characterization of ATR-FTIR spectra requires assigning vibrational peak modes and monitoring relative changes in peak position, width, and intensity. For example, Tsai et al. used ATR-FTIR to qualitatively asses BSA adsorption to PEG-functionalized gold nanoparticles. The authors monitored observed peak broadening and decreased intensity the amide I region, which they attributed to a loss of secondary structure (α-helices). Quantitative characterization was achieved by monitoring changes in peak intensity in the amide II region as a function of BSA concentration introduced to the flow cell housing the gold nanoparticles. This vibrational region is relatively insensitive to protein secondary structure and can be used to quantify total protein adsorption. Using the Langmuir adsorption model, the authors used these data to determine the adsorption density and surface binding constant of BSA on PEG-functionalized gold nanoparticles. They found fairly good agreement between these parameters as determined by ATR-FTIR, a fluorescence assay of adsorbed protein concentration, and electrospraydifferential mobility analysis (ES-DMA, vide infra).27 Quantitative analysis of protein adsorption to or desorption from nanomaterials can, however, be significantly complicated by spectral artifacts due to, e.g., experimental drift, introduced especially during blank subtraction.83 In addition, distinguishing spectral intensity changes resulting from protein exchange with the nanomaterial surface versus protein conformational changes can be difficult, as has been described in the case of human plasma protein adsorption to super-paramagnetic iron oxide nanoparticles.87 For these reasons, ATR-FTIR is best used in combination with other techniques to probe molecular adsorption to nanomaterial surfaces. Surface enhanced Raman spectroscopy (SERS) is another surface-sensitive method capable of probing biomolecule interactions with nanomaterials. SERS relies on the significant enhancement of Raman scattering intensity that occurs within a nanometer-length region of a plasmonically active surface (like a noble metal nanoparticle). Consequently, this technique is inherently limited in scope and has been infrequently employed to characterize biomolecule-nanomaterial interactions. However, by exploiting the fact that surface enhancement occurs over a nanometer length scale, SERS can be used to probe biomolecule binding distance from a plasmonically active surface. For example, Treuel et al. observed that Raman features of proteins adsorbed to citrate-functionalized silver and gold nanoparticle surfaces diminished when the capping agent was replaced with a thicker polymer coating.88
ACS Paragon Plus Environment
Page 12 of 55
Page 13 of 55
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Circular Dichroism. Protein adsorption to nanomaterial surfaces commonly results in a loss of secondary structure. Consequently, circular dichroism (CD) spectroscopy can be applied to quantify protein adsorption to nanomaterial surfaces by monitoring loss of protein secondary structure in the bulk solution.89 A theoretical treatment of the effects of protein-to-nanoparticle adsorption on protein secondary structure has been developed by Li et al.90. The authors used coarse-grained molecular dynamics simulations to model the interaction of apolipoprotein A-1, the main component of highdensity lipoprotein, with silver nanoparticles, and validated these simulations against direct CD measurements.90 In both their simulations and experiments, the authors observed destruction of protein α-helices and increased β-sheet character following interaction with nanoparticles. Gagner et al. observed loss of both α-helix and β-sheet secondary structure in both lysozyme and α-chymotrypsin following their adsorption to gold nanospheres and nanorods.91 Lysozyme-nanoparticle conjugates showed greater losses in secondary structure (and protein activity, measured in parallel using enzymatic assays) than α-chymotrypsin-nanoparticle conjugates. Rotello et al. also observed loss of α-helix structure in BSA upon interaction with gold nanorods funtionalized with amine-terminated polyethylene glycol.92 In contrast to these studies, Wang et al. observed that weak, reversible binding of Subtilisin Carlsberg, an enzyme representative of serine proteases, to silica nanoparticles resulted in no loss of secondary structure.93 However, the authors observed altered near-UV CD spectra following enzyme interaction with nanoparticles, suggesting altered tertiary structure. Despite being weak, these enzymenanoparticle interactions decreased the catalytic efficiency of the enzyme, with smaller nanoparticles inducing a greater reduction than larger nanoparticles. Nuclear magnetic resonance spectroscopy (NMR) is based on a more general detection principle than are the previously mentioned techniques, which complicates its use in complex biological matrices (e.g., those containing multiple amino acids or proteins). However, its high chemical sensitivity and versatile detection platform, including the ability to be combined with measurement of diffusion coefficients, makes it a valuable method for probing biomolecule interactions with nanomaterial surfaces. Solid-state NMR can provide detailed information about the chemical environment at nanoparticle surfaces. For example, solid-state NMR has been used to characterize the orientation of L-cysteine on gold nanoparticle surfaces based on measurements of discrete 1H, 13C, and 15N resonances.94,95 Both broad and sharp 1H resonances were observed, with the former attributed to cysteine bound close to the nanoparticle surface and the latter attributed to cysteine present in an outer layer not chemisorbed to the nanoparticle surface.95 Solution-based NMR also has been used to characterize adsorbed species. Hellstrand et al. used 1H/1D NMR to identify adsorbed lipids that had been extracted (into chloroform) from lipoproteins bound to the surface of copolymer nanoparticles.72 Giri et al. used 1D and 2D (DOSY) 1H NMR to observe the strength of interaction between human serum albumin (HSA) and poly(amindoamine) dendrimers of various generations and with various terminal groups (e.g., NH2 vs. OH).96 Their experiments revealed overall weak interactions between the dendrimers and HSA, indicated by small chemical shifts and changes in diffusion constants. The authors also employed saturation transfer difference (STD) NMR to identify protein domains displaying the strongest interactions with dendrimers. In STD NMR, short RF pulses are used to saturate host molecule (e.g., protein) resonances without affecting other molecules in solution.97 Guest molecules that bind to the host experience saturation through intermolecular saturation transfer, while molecules that do not bind remain unsaturated. The difference spectrum resulting from pre- and post-saturation NMR spectra reveals the chemical environment of only the
ACS Paragon Plus Environment
Analytical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
bound guest molecules. Protons closer to the dendrimer core were more saturated (as a result of saturation transfer from HSA) than those farther from the core, indicating stronger interactions between HSA and inner dendrimer protons than outer protons.96 Mass Spectrometry provides a more sensitive and quantitative means to identify biomolecules adsorbed to nanomaterials than does NMR spectroscopy. Multiple mass spectrometry platforms have been applied to characterize molecules adsorbed to nanomaterial surface; these vary in their sample preparation requirements (e.g., need to separate sample components prior to mass spectral analysis), which subsequently determines their time resolution. Sisco et al. used 2D gel electrophoresis and LC-MS/MS to identify proteins adsorbed to polymerwrapped gold nanorods.98 Nanorods terminated with negatively or positively charged head groups or polyethylene glycol were incubated in rat cardiac fibroblast cell culture media, allowing proteins to adsorb to the nanorod surface overnight. Adsorbed proteins were eluted by increasing the solution salt concentration, separated by 2D gel electrophoresis, and identified using LC-MS/MS. Adsorption of biglycan, a protein that influences cellular interactions with the extracellular matrix, was found to correlate with the impact of nanorod exposure on fibroblast contractile ability: negatively and positively charged nanorods adsorbed biglycan and decreased contractile ability, while polyethylene glycolfunctionalized nanorods resisted biglycan adsorption and no effect on contractile ability. Similar methods based on protein separation by 2D gel electrophoresis and detection by LC-MS have been employed to characterize the nanoparticle protein corona in other studies.99,77,100–102 However, these methods generally lack the time-resolution required to observe the dynamics of protein corona evolution; unbound proteins are typically separated from nanoparticles and nanoparticle-protein complexes via centrifugation or size-exclusion chromatography, and the length of this separation procedure determines the time-resolution of the technique. Higher time-resolution has be obtained by centrifuging nanoparticles and proteins through a sucrose cushion to isolate nanoparticle-protein complexes.103 Subsequently, label-free LC-MS proteomics was used to qualitatively and quantitatively characterize protein adsorption to silica and polystyrene nanoparticles. The high time-resolution of this approach allowed for observation of rapid (
